J. Synchrotron Rad. (2001). 8, 625±627 # 2001 International Union of Crystallography � Printed in Great Britain ± all rights reserved 625

catalysis

The architecture of Mg(II) centres inMAPO-36 solid acid catalysts

Gopinathan Sankar, a* David Gleeson, a C. RichardA. Catlow, a John Meurig Thomas a and Andrew D.Smith b

aDavy Faraday Research Laboratory, The Royal Institutionof Great Britain, 21 Albemarle Street, London, W1S 4BS,Great Britain, bCCRLC Daresbury Laboratory, Warrington,Cheshire WA4 4AD, Great Britain.Email: sankar@ri.ac.uk

The local structure around Mg2+ ions of a Magnesium substitutedaluminophosphate, with the ATS structure (MgAPO-36,Mg/P=0.08), in the as-prepared and calcined state has beeninvestigated by Mg K-edge XAS spectroscopy. High qualityXAS data were collected using the solid-state fluorescencedetector. Mg2+ is found to replace tetrahedrally co-ordinated Al3+

in the as-prepared state and remained intact even aftercalcination, thus yielding a highly active, solid acid catalyst.

Keywords: Magnesium Aluminophosphate, Acid catalysis,XAS

1. Introduction.

The substitution of small amounts of divalent metal ions in placeof Al3+ in microporous aluminophosphates can be utilised tointroduce redox and/or acidic centres, which are uniformlydistributed throughout the framework producing powerful andshape selective heterogeneous catalysts. The catalytic propertiesdepend crucially on the nature of substitution. When these ionsare present in the framework sites, upon removal of the chargecompensating organic template molecules, a loosely boundproton is attached to the bridging oxygen, to charge balance theoverall negative charge due to the presence of divalent ions.Thus, it is important to determine the nature of substitution of thedivalent ions to evaluate the ability of these materials for catalyticapplications. X-ray diffraction, although essential in determiningthe overall structure to understand the shape selective propertiesof these solids, is not completely useful, since the amount ofdivalent ions incorporated is usually low (less that 10 percent)and they are distributed randomly in the structure withoutpossessing any long-range order. X-ray absorption spectroscopy,on the other hand, provides the required structural information,since this technique does not depend on the long-range order.This technique has been successfully used in various transitionmetal substituted materials [Sankar and Thomas, 1999, Sankar,Thomas and Catlow, 2000]. However, recording high-qualityEXAFS data for magnesium containing solids have been aproblem due to the low concentration of the metal ions and thedata collected by electron yield method was not satisfactory; insome cases it was restricted only to the XANES part of the XASdata [See for example, Howe et al, 2000, Presented in thisconference]. It is therefore, our aim to use Mg K-edge X-rayabsorption spectroscopy to determine the charge state as well asthe local coordination environment of magnesium ions in theMg2+ substituted aluminophosphate, MgAlPO-36 material. Withthe recent developments at Daresbury Laboratories in both the

measurement of Mg K-edge EXAFS data by the fluorescencemethod as well as the introduction of a recently developedheating system, it was possible to obtain high-quality Mg K-edgeXAS data to determine the local structure of the Mg2+ ions[Smithet al 1998]. This study clearly shows that Mg2+ ions are presentin tetrahedral environment in both the as-prepared and calcinedstate.

2. Experimental

2.1 Synthesis.

Monophasic MgAPO-36 was prepared using the proceduredescribed in Wright et al (1992). MgAl2O4 was prepared usingthe standard nitrate decomposition route, and MgO was obtainedfrom Aldrich.

2.2 Characterisation.

Mg K-edge EXAFS measurements were carried out at Station 3.4of the Daresbury Laboratory, which operates at 2 GeV with atypical current of 150-250mA. Station 3.4, was equipped with adouble crystal beryl monochromator. Multiple scans collectingboth the total electron and fluorescence yields (using a singleelement solid state detector) were obtained for each of thematerials. The samples were mixed with graphite, pressed to forma self supporting wafer and then mounted onto the sample holderusing double-sided graphite tape. We used a Beryl double crystalmonochromator in these experiments, as although YB66(400)crystals are often preferred for XAS measurements below 2 keV,the presence of two pronounced features in the Mg EXAFSspectra at 1385 and 1437 eV, attributed to a spurious reflectionsfrom the (600) plane (Tanaka et al, 1997) makes the collection ofgood quality EXAFS data extremely difficult (Smith et al, 1998).Furthermore, due to the low concentration of Mg in the MAPO-36 material, the use of the total electron yield method forcollection of EXAFS gives spectra with a very poor signal tonoise ratio. Thus, we used the fluorescence data, which gave farsuperior XAS data quality.Calibration, and background subtraction of the data wasundertaken using the programs EXCALIB and EXBROOK,available at the Daresbury Laboratories, U.K. Final data fittingwas carried out using the EXCURV98 program. (Binsted et al,2000). Phase-shift calculations were carried out using theground state potential and phase shifts (Barth and Hedin, 1972;Hedin and Lundquist, 1969). EXAFS analysis was undertaken forthe first shell only using the single scattering curved waveapproximation, since the data range was limited due to thepresence of large amounts of aluminium in the sample.XRD patterns were collected using a Siemens D5000diffractometer using Cu Kα radiation. FTIR spectra wererecorded on a Perkin-Elmer 1725X spectrometer using an in-situcell (Barrett et al, 1996). Elemental analyses were obtained byICP analysis (Kingston Analytical Services).

3. Results and Discussion

The X-ray diffraction patterns of the materials studied confirmedthe presence of monophasic and highly crystalline MgAPO-36.Furthermore, no collapse of the microporous framework of theMgAPO-36 sample was observed upon calcination (to remove theorganic template molecule). Chemical analyses confirmed a

catalysis

626 Gopinathan Sankar et al. J. Synchrotron Rad. (2001). 8, 625±627

Mg/Al ratio of 1:2 for the magnesium aluminate standard, and aMg:P ratio of 0.08 : 1 ratio for the MgAPO-36 material.First we discuss the EXAFS results of the standard materials andsubsequently the MgAlPO-36 catalysts.

3.1 XAS Studies of MgAl2O4 and MgO standards..

MgO and MgAl2O4 were utilised as model compounds todetermine the structural parameters for Mg2+ in an octahedral andtetrahedral environment, respectively. Spinel structures are oftenused as model compounds for metal substituted alumino-phosphates [Barrett et al, 1996]. However, MgAl2O4, as opposedto MgCr2O4, was specifically chosen in this instance, because notonly it has been shown that this material contains Mg2+, entirely,in tetrahedral co-ordination (Sasaki et al, 1979), but also, like theMgAPO-36 materials it also contains aluminium. This is veryimportant because the Al-K edge is only some 250 eV above theMg-K edge and thus substantially reduces the data range of theMg EXAFS leading to significant truncation effects and errors inthe derived EXAFS parameters. A total of four scans werecollected and averaged for each sample. The derived structuralparameters for these materials are given in Table 1. Thecorresponding EXAFS plots and their Fourier transforms aregiven in Figures 1 and 2 for MgO and MgAl2O4, respectively.The derived Mg-O distances (see Table 1) show good agreementwith those obtained from the crystal structures of these materials(2.109 Å and 1.922 Å for MgO and MgAl2O4 respectively)

-15

-10

-5

0

5

10

15

20

25

3 4 5 6 7

k / Å-1

( χ)k

.k3

0

20

40

60

80

0 2 4 6 8 10

r / Å

FT

mag

nitu

de

Figure 1Mg-K edge EXAFS fit (top) and associated Fourier transform (bottom)for MgO standard. [Experimental data (), theory (- -)].

Standard Oxygen environment R(Mg-O) / Å 2σ2 / Å-1

MgAl2O4 Tetrahedral 1.958 0.010

MgO Octahedral 2.128 0.024

Table 1Structural parameters of tetrahedral (MgAl2O4) and octahedral (MgO)standards obtained from Mg-K edge EXAFS data. R(Mg-O) is the Mg-Ointer-atomic distance, and 2σ2 the Debye - Waller factor.

-15

-10

-5

0

5

10

15

20

3 4 5 6 7

k / Å-1

( χ)k

.k3

0

10

20

30

40

50

60

0 2 4 6 8 10 12

r / Å

FT

Mag

nitu

de

Figure 2Mg-K edge EXAFS fit (top) and associated Fourier transform (bottom)MgAl2O4 standard. [Experimental data (), theory (- -)]

3.2 XAS Studies of MAPO-36 catalysts.

Mg K-edge EXAFS data were analysed using a procedure similarto that used for the standards using an identical k-range, andwithout Fourier filtering the data. It must be noted, however, that these materials readily adsorbwater that can only be desorbed at high temperatures. Thus,unless the catalyst is dehydrated in-situ, prior to recording theXAS measurement (as was done in this case), a third possible Mgenvironment (apart from the regular tetrahedral and octahedralcoordination geometries) consisting of four oxygen’s with bonddistances typical of those seen for the tetrahedral site, and twoextra oxygen at ca 2.4 Å attributable to the loosely bound watermolecules, should be considered as a possible model for theanalysis of the EXAFS data.Refinement of the values of the inter-atomic (Mg-O) distances,the Debye-Waller factors and Eo were carried out starting fromeach of the possible three models to yield the best fit. The derivedstructural parameters for the best fit for our MgAPO-36 catalyst(Mg/Al 0.08) before and after calcination are given in Table 2.The corresponding EXAFS plots and their Fourier transforms are

J. Synchrotron Rad. (2001). 8, 625±627 Received 31 July 2000 � Accepted 2 January 2001 627

catalysis

-15

-10

-5

0

5

10

2.5 3.5 4.5 5.5 6.5 7.5

k / Å-1

( χ)k

.k3

Figure 3Mg-K edge EXAFS fit (top) and associated Fourier transform (bottom)for MAPO-36 catalyst before calcination. [Experimental data (), theory(- -)], (inset) Schematic model of Mg2+ site present in as-synthesisedcatalyst

given in Figures 3 and 4 for MgAPO-36 before and aftercalcination, respectively. The derived structural parameters forthe MgAPO-36 catalyst before calcination, indicates that thereare four oxygen’s at 1.935 Å in the first coordination spherearound the magnesium. Comparison of the Mg-O distances withthose obtained for the standard materials indicate that magnesiumis present in the framework, in place of a tetrahedrally co-ordinated Al3+ ion. After calcination of the MAPO-36 catalyst, itcan be seen that the Mg-O distances are almost identical to thosein the as prepared material. Therefore, the magnesium ions arelikely to be present in the framework after calcination. Thus asthe x-ray diffraction pattern indicates that the long range structureis preserved on calcination, and EXAFS indicates that the Mg2+ isretained in a tetrahedral framework position it is expected thatthis material will be a good catalyst. Furthermore, although it wasshown in the case of cobalt substituted CoAlPO’s [Barrett et al1996] that upon introduction of acidic proton, the tetrahedral siteis in a distorted environment, a similar detailed study is notpossible in the case of magnesium, since the available data rangeis rather restricted due to the presence of aluminium, asmentioned earlier. EXAFS analysis starting from the hydratedmodel, gave non sensible values for the Debye-Waller factor ofthe oxygen’s attributed to water and a poorer fit to theexperimental data than that obtained for the dehydrated model.We have shown that high quality Mg-K edge EXAFS data can beobtained from a material containing low concentrations of Mgusing the solid state soft-energy fluorescence detector on station3.4 of the Daresbury Laboratories, U.K. The use of MgO andMgAl2O4 standards has allowed us to determine that Mg2+

substitutes directly into the aluminophosphate lattice in place oftetrahedral Al3+ in MAPO36. This geometry is retained upon

calcination (to remove the occluded template and activate thecatalyst) yielding a material which is a good solid-acid catalyst.

-10

-5

0

5

10

2.5 3.5 4.5 5.5 6.5 7.5

k / Å-1

( χ)k

.k3

Figure 4Mg-K edge EXAFS fit (top) and associated Fourier transform (bottom)for MAPO-36 catalyst after calcination. [Experimental data (), theory(- -)], (inset) Schematic model of Mg2+ site present in calcined catalyst.

Material N and type R / Å 2σ2 / Å-1

MAPO-36 4 x O 1.935 0.012(As prepared)

MAPO-36 4 x O 1.943 0.008(Calcined)

Table 2Structural parameters of MAPO-36 catalysts before and after calcinationobtained from Mg-K edge EXAFS data. N is the coordination number, Ris the inter-atomic distance, and 2σ2 the Debye - Waller factor. Typicalerrors: N ± 10%; 2σ2 ± 10%; R ± 0.02 Å

The authors thank EPSRC for the financial support of thisproject, CCRLC and Dr N. Binsted for the EXAFS data analysissoftware support. G.S. thanks the Leverhulme Trust for a researchfellowship.

ReferencesBarret, P. A., Catlow, C. R. A., Sankar. G., Thomas, J. M. (1996).J.Phys.Chem. 100, 8977-8985.Barth, U. v., Hedin, L., (1972), J.Phys.C., 5, 1972-Binsted, N., Gurman, S. J., Campbell, J. W., Stephenson, P. C., (2000),EXCURVE, Daresbury Laboratory program.Hedin, L., Lundquist, S., (1969), Solid State Phys., 23, 1Sankar. G, Thomas. J.M., (2000) Topics in Catalysis, 8, 1Sankar. G, Thomas. J.M., Catlow. C.R.A. (2000),Topics in Cat.10,265Sasaki, S., Fujino, K., Takeuchi, Y., (1979), Proc.Japan.Acad., 55, 43.Smith, A. D., Cowie, B. C., Sankar, G., Thomas, J. M., (1998),J. Synchrotron Rad., 5, 716-718.Tanaka, T., Aizawa, T., Rowen, M., Rek, Z. U., Kitajima, Y., Hogashi,I., Wong, J., & Ishizawa, Y., (1997). J.Appl.Crsyt. 30, 87-91.Wright P.A., Natarajan, S., Thomas, J.M., Bell, R.G., Gaiboyes, P.L.,Jones, R.H. Chen, J.S.,(1992), Angew.Chem.Int.Edn.Engl., 31, 1472

0

10

20

30

40

50

0 2 4 6 8 10 12

r / Å

FT

Mag

nitu

de

P5+ Mg2+

Mg2+P5+

0

20

40

60

0 2 4 6 8 10

r / Å

FT

Mag

nitu

de